MEGA-Bench Task - Red Teaming Captcha: leaderboard

Metric: Task Score (%). Source: huggingface.co. 44 models tracked.

Top models

#ModelScore
1Qwen 2 VL 2B12.1
2InternVL2.5-2B11.1
3Gemma 3 27B (IT)10
4GPT-4o10
5Gemma 3 12B (IT)10
6Gemma 3 4B (IT)10
7InternVL3-78B10
8InternVL3-8B10
9InternVL2-8B10
10Qwen 2 VL 72B10
11Qwen 2 VL 7B10
12Claude 3.5 Sonnet (20241022)10
13InternVL3-38B10
14Aquila-VL-2B10
15GPT-4o Mini10

Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-red-teaming-captcha · How It Works · Data refreshed daily, snapshot 2026-09-05.